Multivehicle Task Assignment Based on Collaborative Neurodynamic Optimization With Discrete Hopfield Networks.

Area of Science:

• Robotics and Control Systems
• Operations Research
• Computational Neuroscience

Background:

• Multivehicle task assignment (TA) is a complex combinatorial optimization problem.
• Existing methods may struggle with dynamic constraints and large-scale problems.
• Efficient task allocation is crucial for operational efficiency in multi-robot systems.

Purpose of the Study:

• To develop a novel neurodynamics-based approach for multivehicle task assignments.
• To reformulate the TA problem into a solvable Quadratic Unconstrained Binary Optimization (QUBO) format.
• To demonstrate the effectiveness of the proposed Collaborative Neurodynamic Optimization (CNO) algorithm.

Main Methods:

• Reformulation of the TA problem as a QUBO problem incorporating utility and penalty functions.
• Development of a TA algorithm within the Collaborative Neurodynamic Optimization (CNO) framework.
• Utilizing a population of Discrete Hopfield Networks (DHNs) to solve the QUBO problem.

Main Results:

• The proposed CNO-based TA algorithm successfully solved the reformulated QUBO problem.
• Experimental results demonstrated superior performance compared to existing methods in various scenarios.
• The approach effectively handled load capacity and cooperation constraints.

Conclusions:

• The neurodynamics-based TA approach offers an effective solution for multivehicle task assignments.
• CNO provides a robust framework for solving complex optimization problems in multi-robot coordination.
• The proposed method shows significant potential for real-world applications in logistics and autonomous systems.